Drew DeVault ( @sir )has a new blog entry making a case for the Web being dead.

"We have no recourse left to preserve the web. This is why I'm throwing my weight behind Gemini, a protocol which is much simpler than the web, and which you can implement yourself in a weekend. Forget about the web, it's a lost cause. Let's move on."



#Gemini #SmallInternet #Blog

@dublinux @sir I hear and feel that frustration with what client-side web technology has become. but hypertext is really still not that complicated. you can write a web server in less than twenty lines of most popular programming languages (albeit not a secure or scalable one, and you'll rely on another http implementation), but if simplicity is what you want, look to the #indieweb movement for accessible, hackable tools.

@travisfw @dublinux @sir Also: The Web can be *almost* as simple as Gemini (depending on how much you want to implement) if only sites would stop relying on JavaScript (and less importantly CSS).

Which I find a majority of websites do, it's just the popular sites which are a problem!

That's my experience anyways!

@alcinnz @dublinux @sir yeah JS always seemed like a mistake. from my first exposure to it in 2001.


@travisfw @dublinux @sir Yeah, now the question is: How do we put the genie back into the bottle?

@alcinnz @travisfw @dublinux @sir This is where Stallman (controversial I know) is right inn being annoying, if you go browse a site with firefox with no-script installed and this site shows you a blank page most users would say no-script makes the web unusable (heard it on a podcast one time) but the correct action would be to notify the author that their page is broken, because it is... a website should be written in HTML according to W3C and everything is else is adidtional!

@gassahara I use noscript and it's really frustrating how many things are completely broken with it

For interactive sites like on fedi it's justifiable but at least they expose nice apis for it, but so many news sites just *don't work*

and even things like birdsite make you wait several seconds before prompting you to redirect to "legacy twitter", why does that make any sense?

@lunch oh, on birdsite is really anoying, since you can view images and *some* replies but only after they keep repeating you: hey! this is weird, are you sure you're doing it right? =)
and the legacy version is not their legacy version, i remember that site and it had all replies, and images, they don't show you that now, it's really cumbersome =)
But No-Script is liberating, you don't want my way? you don't have my eyes :)

@lunch @gassahara I use nitter.net to follow links to birdsite, as it does not require JavaScript.

@sankakujin @lunch Thank you, will try it, had trouble testing right now, surely after some tinkering i'll get it =)

@alcinnz @travisfw @dublinux @sir i don't think the genie goes back in.

but. imo the web can make sense again. webcomponents (specifically custom elements) is a huge step, to recursively using our patterns. this is hypertext- it makes sense that we ought to have new symbols, new hypertextual elements arise. that fit & work together like the existing elements.

the more we can do to move state from JS onto the web itself, the better. at the moment, it's kind of a philosophical jump without purpose, but long term, i think it can lead to advanced interoperation, to new heights of scriptability, to new forms of hypertextuality.

@alcinnz @travisfw @dublinux @sir the genie remains, but, like, the world can move on. we can get good again, & create a better web, create interesting hypertext, that is true to the promise, where (unlike code) what is there is visible & known & hypertext.

and it can & will be so much better, technically, & over time, for the users. that no one will want to do this nasty vicious deeply frameworked stack stuff. the stack can disperse, back into the web, once we start creating elements to represent all these things we've hidden away inside JS.

@jauntywunderkind420 @alcinnz @dublinux @sir webcomponents seems like a good step forward for JS, but a step backward for HTML. looking at a couple tutorials, it looks like the display of the custom HTML tag would be relegated to a procedural JS class. that's not a machine readable ontology. it may make HTML more concise, but at the cost of making its meaning more obscure, and enabling obfuscation.

@travisfw there are works underway to create Declarative Custom Elements in html. it's a long way off, but there are a lot of interested parties.

even in the current form, with the component having to be authored in JS (perhaps with lit-html inside, to author the HTML), while it may not be as declarative as we'd want, it still lets the surface object be representable in html, whereas right now all the JS machinery is implicit, is inside the JS. letting these JS machines expose themselves & be composed is a major upgrade over the 100% imperative construction tools we use today to build apps.

Sign in to participate in the conversation

For people who care about, support, or build Free, Libre, and Open Source Software (FLOSS).